Articles Trophic Interaction Cycles in Tundra Ecosystems and the Impact of Climate Change ROLF A. IMS AND EVA FUGLEI While population cycles are geographically widespread, it is on arctic tendra that such cycles appear to be most influential for the functioning of the whole ecosystem. We give an overview of tundra species that exhibit population cycles and describe what are currently believed to be the causal mechanisms. Population cycles most likely originate from trophic interactions within the plant-based tundra food web, where lemmings, either as prey for carnivores or as consumers of plants, play the key role. The predominance of trophic interaction cycles af northern latitudes is ultimately related to climate and such cycles should therefore be vulnerable to climate change. Recent evidence indicates that changes have already taken place in the dynamics of some key herbivores and their predators, consistent with the expected impacts of climate change. There is a strong need for large-scale integrated monitoring and research efforts to further document such changes and their ecosystem consequences. Keywords: arctic tundra, climate change. ecosystem functioning, food web dynamics, lemmings In this article we provide an overview of what is known about cyclic dynamics in terrestrial arctic ecosystems (i... tundra). First, we take a species-oriented view and describe tundra species exhibiting population cycles. Second, we place these species in an ecosystem context by outlining the basic architecture of the plant-based tundra food web and the types of interactions taking place within this web. We then show how cycles can be a product of trophic interactions by reviewing the most plausible theories and recent empirical evidence. Finally, we examine the role of arctic climate in these interaction cycles and end with a discussion of how climate change may act to alter them and what the wider conse- quences of such changes may be. ife on the arctic tundra is subject to dramatic year- Ito-year variation in terms of bioproduction. In some years wildlife populations flourish, while in others the tun- dra appears remarkably devoid of wildlife. Although indige- nous people and early explorers have always been aware of the violent booms and busts in arctic wildlife, it was not until the English ecologist Charles Elton (1924) started to examine sta- tistics on fur-bearing animals that these multiannual fluctu- ations were found to follow a cyclic pattern. Elton recognized that there were conspicuous peaks in the number of arctic fox skins exported from arctic Canada every 4 years, and he found a similar cyclicity in the Norwegian zoologist Robert Collett's compilation of records on "lemming years" in Nor- way (Lindström et al. 2001). Today, the literature is consid- erably broader: Many thousands of scientific papers on population cycles have been published in the 80 years after El- ton's discovery. The phenomenon is not restricted to arctic species, although it is definitely most common in northern areas (Kendall et al. 1998). Moreover, it is on the arctic tun- dra that population cycles seem to be most intertwined with the functioning of the whole ecosystem. The important ecosystem consequences of population cycles were high- lighted three decades ago during the International Biological Program (e.g., Batzliet al. 1980), but in recent years this per- spective has drawn less attention. The recent realization that climate change will affect arctic ecosystems severely, and that altered cyclic dynamics in tundra species are likely to be in- volved (Callaghan et al. 2004), calls for a renewed focus on the role of such cycles in the Arctic. Arctic species with cyclic population dynamics For laypeople, population cycles are perhaps most conspic- uous in the two species treated in Elton (1924): the arctic fox and the lemming. The cycles in the population of the arctic fox-the most valuable furbearer on the tundra-were, and to some extent still are, influential in the economy of arctic communities. The lemming cycle, on the other hand, repre- sents the most pronounced fluctuations in terms of biomass. Rolf Aims (e-mail: rolfimib.it.no) is a professor of ecology in the Department of Biology, University of Troms, N.9037 Troms, Norway, Ev Fuglei (e-mail: eva.fugleipolar.no) is a research biologist at the Norwegian Polar Institute, Polar Environment Centre, N-9296 Troms, Norway 2005 American Institute of Biological Sciences April 2005 / Vol. 5 No.. BioScience 311
Gyrfalcon Red and arctic Parasiticaeger Musk ox Carbou shorebirds Hante Grasses and Figure 4. Outline of a typical high-arctic plant-based food web. Components of the food web involved in lemming population or production cycles are in bold frames and linked with thick arrows. Thick, solid lines indicate direct relationships with lem- ming cycles, while dashed lines indicate indirect relationships (i.e., alternative prey mechanisms). Modified from Krebs and colleagues (2003). mings possess key species attributes (i.e., they are likely to in driven plant production cycle, which is simply converted teract strongly and dynamically with many components of the into herbivore population cycles. Alternatively, the cycles food web; figure 4). For this reason, we center our discussion may be the outcome of plant-herbivore interactions involv- of possible cycle-generating mechanisms on lemmings and ing grazing-induced changes in plant quality, or they may re- their trophic interactions with plants and predators. sult from changes in plant quantity. Internally driven plant production cycles. The idea of an How are interaction cycles generated? internally driven plant production cycle stemmed from the The origin of lemming and vole population cycles has been observation that good production years in tundra plants co- sought ever since Elton's 1924 paper, and some 30 to 40 hy incided with lemming peak years even when plants were potheses have been put forward. Several general overviews of protected within exclosures (and thus were not subject to graz- this research on population cycles in small mammals are ing) (Laine and Henttonen 1983). Production cycles in peren- available (e.g., Stenseth and Ims 1993, Korpimäki and Krebs nial plants can be generated if energy reserves must be 1996, Turchin 2003, Korpimäki et al. 2004). Here we restrict accumulated over several years to attain thresholds for suc- our focus to mechanisms that may underlie lemming and vole cessful seed production. Synchronization within and be- cycles in the context of arctic food webs. Indeed, the current tween different plant species will then be brought about by view is that such population cycles cannot be understood un climatic variation (Laine and Henttonen 1983). This mech- less they are viewed as an integral part of the food web (Berry anism has been mathematically validated and is now thought man 2002, Turchin 2003). to underlie the general phenomenon of mast production in many perennial plants (eg, Satake and Iwasa 2002). Even Plant production cycles and plant-herbivore interactions. though plant production cycles can be expected on theoret- There are three ways by which plants may be involved in the ical grounds, empirical evidence for them in arctic plants is generation of interaction cycles. There may be an internally poor and partly contradictory (e.g., Oksanen and Ericson April 2005 / Vol. 55 No.4. BioScience 315